How docs get into the tech game

Dr. Mark Frisse, MS, MBA, is a Professor of Biomedical Informatics and Vice Chair of Business Development at the Vanderbilt University School of Medicine. His diverse background includes work in computer science, management, clinical medicine, consulting, and technology research.

Most notably, Dr. Frisse was part of a team that aimed to develop early e-prescribing standards while working for Express Scripts, setting up an organization called RxHub in an effort to develop universal norms for electronic prescriptions.

In 2004 while working through Vanderbilt, Dr. Frisse helped to establish a health information exchange (HIE) organization in Memphis, TN, that linked all hospitals in the region to a single system, in an effort to allow all digital information to be easily accessed by every emergency department and health system in the greater Memphis area.

I’ve heard you’ve had an interesting journey. Can you tell me a little about your background?

I’ve been the Forrest Gump of medical informatics in that I’ve been in the middle of some things. I was in the genome project as a researcher and watched people meet with James Watson, of Watson and Crick. You can find a YouTube video of me making a carefully crafted public apology to Tim Berners-Lee because it was, I think 1991, at an early hypertext meeting I read one of his papers and kind of said it wouldn’t scale up. At the time, it really wouldn’t. I can defend that decision. But see, I was one of the first people to do hypertext, in 1987 as part of my master’s degree. The idea was you typed words into something called “the window” and clicked on these things called “hypertext links.” And it was basically doing the same thing Google does today. I just wasn’t smart enough to see the potential!

One of my employees, a kid named Scott Hassan, ended up going out to help write the Google search engine as one of the founders. He’s a billionaire now.

So, my course has been atypical. I’ve worked with technology companies. I worked as a consultant for a group that helped people pick EHRs, and I’ve been an academic teacher – that’s my primary goal now.

I also worked under the governor of Tennessee on the ground, trying to get disparate healthcare organizations to collaborate – and this was fairly early on. We were one of the first health information exchange groups. This was before ONC was even formed.

It isn’t all the time that I talk to a tech-savvy doctor. What motivated you to combine the two for your career?

In the 1980s, as a chief resident at Barnes Hospital, I was responsible for managing the charity budget, and I was appalled by both healthcare costs and the disconnect between what physicians knew of cost and what was really going on. I really can remember very well the faces of all the people who could not get care because of the cost – and we had a very generous charity budget.

At the same time I was using a word processor, because I was writing a book. And so I said, “Holy cow, you could somehow put information on computers.” That seemed important.

That lead me to go back to Stanford University in 1985-87, and I got a master’s degree in informatics there, right at the beginning of that program. At Stanford I was a trainee with people like Eric Horvitz from Microsoft and a guy named Marty Chavez, who is presently the CIO of Goldman Sachs.

And teaching is now your primary focus?

Yeah, I’m a teacher now at Vanderbilt University. At Washington University – I was there for medical school until I left after 25 years – I was a full professor and an associate dean. While there, I met the CEO of Express Scripts who wanted to build a website for 60 million Express Scripts customers. So, I left Washington University and joined Express – and I learned quickly that, just because you’re a very smart doctor doesn’t mean you’re an effective manager or a good businessman. I developed a very, very healthy respect for all of the different disciplines that are required to make our healthcare system work.

Is this where RxHub comes in? What was its role in developing a standard for e-prescriptions?

RxHub was in 1999 and 2000. We worked with, at the time, MedCo Health Solutions and others. Teams of four of us from each company would meet secretly – along with a bunch of antitrust lawyers – and we’d find common standards, so when we built e-prescribing devices, people could find out what’s in their formula. That was it. We were kind of ahead of the curve on that.

Was there a need for e-prescribing standards back then? I can’t imagine there were all that many electronic prescriptions.

It was the CEO of Express Scripts who understood even in 1999 that if we do not standardize these things somehow – and our systems have to talk to every one of these things – we’re going to be in trouble. Express Scripts figured out early on the same thing the federal government has just now figured out – we’re going to need a standardized approach to allow people choices for their technologies, or we will have chaos. And that’s what led to all of us getting together to do this.

We had all sorts of rules – you couldn’t show preferred drugs or ads. We created a compact among ourselves to focus on value and clarity, and leave the choices up to the consumers. And I can tell you from working with Express Scripts that we always looked at therapeutic value first, then we looked at which would be cheapest for our customers. We never had a policy about which drug to give our customers. We never did that. And that’s something I’m pretty proud of.

Tell me about your work with a health information exchange. Why is it so hard to share health information between health systems?

The first thing to understand is that form follows function, and our information systems are designed for the behaviors we really honestly want, and sadly the electronic health records we want have evolved more like a “smart book” than what they’re supposed to be, which is a modern communications device.

Many of us were deeply, deeply concerned with the naïve approach to health information exchange that was in the HITECH law, and to this day, it’s a challenge. Why? Because “health information exchange” is not a noun. Granted, the health information exchange I built in Memphis was a noun; it was an organization with an architecture. But “health information exchange” is a verb. It’s the ability to get information from one system to another when you need it, in a secure and honest way. When you focus on the noun, you almost always fail.

One day we’re going to see a change, starting from the peripheries and coming in, that’s going to radically change how we manage healthcare information. How will it be done? Don’t know. But you don’t have to be a genius to predict the future with this stuff – look at the money, look at the trends, look at the technology. I think people will just have to figure it out, and it’s going to be messy. But they’ll figure it out because they have to.

On that note, what’s your view of accountable care models and a move toward value-based care?

I think our entire field has been full of acronyms and hyperbole that have been disappointments. There’s a new buzzword every year that people are using at HIMSS, and they are always disappointments. But if you look at it as a macro trend, it’s inevitable that we have to start looking at care in terms of long-term models. I don’t care what your politics are; I don’t care what you think about markets, whether you want a single-payer system or a completely libertarian system, at the end of the day, in my own view, the only thing that makes sense is to pay over the course of someone’s care, and not for each discreet episode – otherwise you’ll just continue to rack up essentially fee-for-service.

“We have this war of mutually assured destruction between the policies we automate and
the policies we use to regulate – and I believe the overwhelming majority of the mental and computer cycles that we expend in this country are focused on dealing with unnecessary complexity. It’s really that simple.”

So what’s wrong with accountable care organizations? As is always the case – and as was certainly the case with HITECH – we let our rhetoric get ahead of what is possible. For example, HITECH aimed to do 10 to 15 years of culture change in five years by throwing a lot of money at the problem. It’s kind of like turning your oven up to 1,000 degrees instead of 500 degrees so you can bake your pie in half the time – that doesn’t work.

Instead of looking at how we can make systems which address consumer needs and market needs and help consumers make informed choices, HITECH focused all its energy on the electronic health record, which is as it is today as it was in 1975. We basically took the last steam locomotive and, instead of creating an auto industry, they basically told people to buy steam locomotives and they’ll get a lot of money.

It was that top-down push that was the problem. That said, it’s a love-hate thing. I’m so glad we have the EHR adoption we have, even if you adopt the phrase from athenahealth’s Jonathan Bush, “HITECH was cash for clunkers.” I love that phrase he used.

What kind of advice do you give your students – words of wisdom – to make sure they’re working to solve healthcare’s most pressing issues? Assuming they are solvable.

Oh, there are solutions. The first is to ask yourself whether or not a system you’re participating in is really there to create positive value – business and clinical value – or if it’s just there to automate an unnecessary, incredibly complex system. We’re not in the healthcare business, we’re in the complexity business. If we weren’t in the complexity business, why would we have thousands of formularies? Why would we have different prior authorization rules for every form and every plan? Why would we have so many different Medicaid programs?

Ask yourself this: If we didn’t have computers, what kind of tax system would we have? Well, it’s got to be a flat tax or a consumption tax. The only reason we can have all of this crazy tax code is because computers can process it. And of course if computers can make it more and more complicated, that means the tax code and tax software also become more complicated. Healthcare is the same way. We have this war of mutually assured destruction between the policies we automate and the policies we use to regulate – and I believe the overwhelming majority of the mental and computer cycles that we expend in this country are focused on dealing with unnecessary complexity. It’s really that simple.

Once we automate this complexity, and we see it – and we have to make more of these complex systems come together – we face a catastrophe. Look at Healthcare.gov. I think it was arguably the most complicated and largest piece of computer software ever built. And why did it fail? One of the reasons why is because the rules were so complex – you had to know people’s income tax results, what state they were in, what category of plans they were eligible for, what the prices would be. This is like the antithesis of WalMart. This is a system that would make every individual tube of toothpaste on a store shelf cost something different, each managed with their own set of rules. I think a lot of ideas fail because our reimbursement and regulatory system is simply too complex.

The Pragmatic Innovator

Dr. Anil Jain, Senior VP, Chief Medical Officer and Co-Founder, Explorys, founded the cloud-based analytics platform, which was later merged with Watson Health in 2015. Dr. Jain spent most of his career at the Cleveland Clinic, where in his role as Senior Executive Director of IT he spearheaded innovations and research surrounding EMR-related data and population health.

He serves on a number of advisory boards and has authored numerous publications.

Dr. Jain is a Diplomate of the American Board of Internal Medicine and a Fellow of the American College of Physicians.

In your own words, tell me a little bit about yourself and your background.

During my time in med school at Northwestern University, I started really getting involved with understanding information, thinking about the fact that, as a biomedical engineer and having a technology inclination, that healthcare was behind. To put it gently, it was significantly behind when you start looking at other industries.

Shortly after medical school, we go through a process for residency training. I ended up at the Cleveland Clinic, which at the time was, and still remains, one of the most impressive healthcare organizations for primary care and tertiary care. When I arrived for training, we really didn’t have much of an electronic health record. We had some digitalization happening, but it was partially done. And through my training process, I ended up interacting with Martin Harris – who was the Chief Information Officer of the Cleveland Clinic – and I got involved with projects that really started to look at what technology physicians need.

I was training to be an internal medicine doctor, and I provided some perspective to him and to the rest of the IT team at Cleveland Clinic, from the perspective of a practicing physician – someone who undersood technology but also understood medicine. I stayed on at Cleveland Clinic for a chief residency and during my chief residency, I worked on some publications and co-wrote a book with a colleague of mine, and stayed connected to the technology part that was evolving at the Cleveland Clinic, including the selection of Epic as an EMR.

You said the Cleveland Clinic was operating entirely on paper when you arrived there. What type of EMR innovations were you able to push, given the limited technology?

One of the most frustrating experiences that I personally had – and I think many physicians across the country could relate to this – was that, when we would often use the EMR to manage patients, we didn’t see much of a return when it came to the data. We didn’t understand where the data went. Why wasn’t it being used to improve and measure and enhance the way we delivered care? It seemed like it was just a glorified paper medical record turned into an electronic form. It didn’t really do all of the things that I knew from my technology background that it should be able to do.

So, I took it upon myself and got some of my colleagues to really focus on the back end of the EMR, in addition to the front end. I set up a group called eResearch that was built to leverage all of the data that was in our back-end systems and use it for quality improvement, including research and clinical trials.

We incubated eResearch, and it was essentially designed to go after improving quality of clinical research, using all of the data from the EMR in a safe, secure way – HIPAA compliant and all of the things you’d expect from a health system like the Cleveland Clinic. Often I would go around the country talking about what we are doing, and people would say, “How do I get these capabilities?” So, I started thinking about how to commercialize the work that we had done. In 2008, or so, I approached the Cleveland Clinic innovation group, which was really designed to do technology transfers between employees and doctors like myself and figure out a commercialization path.

Is this where Explorys stemmed from?

Explorys spun out of the Cleveland Clinic in October 2009. I stayed on at the Cleveland Clinic as well as being involved with Explorys. And then I left the Cleveland Clinic in June 2011, and then came back a few months later after Explorys got their seed funding from venture capital groups. I came back in and joined the group. At that point, I became a member of the consulting staff at the Cleveland Clinic and a full-time Chief Medical Informatics and Information Officer for Explorys – my job was to bring all of those tasks into one role. And then we really focused on helping healthcare systems around the country to do the same thing. How do they use their data, get it out of the back-end system – liberate their data, if you will – and use it in a standardized way for quality improvement and clinical research?

Did you know at the time this data would be particularly useful for clinical research?

A lot of people describe me as a pragmatic innovator because I try to solve problems that I know are here right now or going to be right around the corner. You have to understand what’s going on in the market, and you have to figure out what part of it do you want to innovate around and what part of it feels like others can do a better job innovating. And so I knew there were challenges in the market – there was a significant problem finding patients to recruit into clinical trials. Even when doctors wanted to get their patients into clinical trials, they didn’t know which patients were the best candidates. That use case was a loud-and-clear use case.

“That is the true promise of technology – to simplify things, not make them more complex. People no longer have to put DVDs into a machine, they can watch a Netflix stream. They no longer have to go to a video store, they can browse a library right on a screen. It’s simpler, not more comple. I think that’s where healthcare needs to go.”

One of the problems that often happens when you get research grants to fuel your work is that you’re basically having to go from research grant to research grant trying to bolster a case. Long ago, having been involved with several projects that died out because the research money dried up, I told myself that the way I was going to innovate in a pragmatic way was to go after a commercialization path. Because if you can convince people there’s a business model, then you don’t become dependent on research grants. You become dependent on a business case and a business model, and a commercial value proposition, if you will. And so that’s the way I approached it. We knew it could be done if we scale it. And that is where I needed to commercialize it and bring in money to get it scaled up, and that’s where Explorys comes from, in October 2009.

The next several years are all about realizing that vision, selling our platform to health systems all across the country. We have about 28 large customer organizations, 400 hospitals that are using our platform in one way or another, and whose data we’ve been able to aggregate to help learn about what’s happening in healthcare.

You mentioned you felt healthcare was behind. Is that still the case, and is data analytics the answer when it comes to helping the industry catch up?

Keep in mind that was 1995 when I came to the Cleveland Clinic, and if you were to ask me if healthcare is still behind, I would still say, “Yes, it is.”

Think about Netflix and Amazon, and the way that they’ve changed their industries. Think of Uber. When I say that healthcare is behind – that’s because it is. It is behind. Part of the reason is because healthcare is not necessarily a system in the way that others might describe it; it’s more of a cottage industry.

Now, to your question of if data analytics is the answer. Well, I’m not sure. I’ll tell you where I see us getting into a little bit of trouble is that we tend to focus on data as a way to bring transparency and equalize knowledge that people need to have to make better decisions, but there’s way too much of it.

I think the answer truly is to take all the data and convert it to useful insights and then use cognitive solutions to deliver those insights to the right people at the right time at the right place. Otherwise, we are still shifting the problem back to the doctor who gets information overload and becomes paralyzed when it comes to deciding what’s the best thing to do.

That is the true promise of technology – to simplify things, not make them more complex. People no longer have to put DVDs into a machine, they can watch a Netflix stream. They no longer have to go to a video store, they can browse a library right on a screen. It’s simpler, not more complex. I think that’s where healthcare needs to go. And this is why I’m truly so excited about Explorys being part of Watson Health, because that cognitive lens that IBM has been nurturing over the past six-plus years is the answer. It is truly the way that we’re going to be able to prevent this information overload from consuming the interactions that docs have with their patients.

What benefits does the merger with Watson Health provide for Explorys?

On April 13 of last year, IBM and their Watson Health unit recognized that Explorys was an absolute leader in bringing data together and doing analytics that helped on the population health and clinical research side – and they bought us to become the pivotal piece of their Watson Health portfolio.

I would say that Watson and Explorys were waiting for each other. Watson, in order for it to understand and to learn and to create hypotheses, and then check those hypotheses and reason why something may be true or false – it needs raw material. It needs to read. It needs to digest information. Explorys data becomes raw material for Watson to ingest, read, and understand what’s happening in the real world – we call that real-world evidence.

On the other hand, Explorys needs Watson. We know much of the data that is buried in healthcare systems today is not structured data entered into specific parts of an EMR or a health information system – it’s buried in the notes that I know I write when I see a patient. It’s buried in free text. How do we make sense of it? At the end of the day, if you want to understand the context and personality behind the person who wrote the materials, you need something cognitive. And what Watson can do for Explorys, in a sense, is it can help us make sense of all that unstructured data that’s out there, and then drive that back into our platform so that our analytics and our innovations in population health and life sciences are as complete as they can be.

We tend to have a lot of biases as docs, and we extend those biases into how we write our notes and how we communicate with a colleague. Watson understands all of that, and it can extract relevant information from it.

The inherent biases we have really shape our ability to innovate and discover. Watson is unencumbered by those things; it’s free to explore and generate hypotheses, discount them quickly, and present back to us the ones that are worth exploring further with our brain. Watson really is just like us. We try, we fail – we learn. When we succeed, we learn from that. Watson is very much the same. That’s why we call it a cognitive agent. It’s not just analytics; it truly is a learning platform.